Lung cancer is a common malignancy among both men and women in the USA. The overall prognosis for this disease remains poor, with 13% survival at 5 years for all stages. However, diagnostic techniques and treatment modalities have improved over the past one to two decades. Moreover, advances in molecular biology and genomics have provided valuable insight into disease susceptibility, which may, in turn, impact tumor recurrence and treatment response. The proposed study, a resubmission in response to PA 98-027 (NCI) takes advantage of a large, well-characterized cohort of lung cancer cases established by the PI in 1992. Detailed information will be collected on a cohort of cases recruited between 1992-1999, and, by the end of the proposed study, each case will have been followed for at least 5 years and as long as 13 years. We will utilize both standard molecular biologic techniques as well as single nucleotide polymorphism (SNP) array hybridization technology to test the following hypotheses: (1) polymorphisms in candidate genes involved in a number of xenobiotic and chemotherapeutic drug metabolism, all cycle, DNA repair and invasion pathways are associated with clinical prognosis in NSCLC patients; (2) gene-exposure (smoking) and gene-gene of these polymorphisms modify clinical prognosis in early- and late-stage NSCLC; and (3) using a genome-wide loss of heterozygosity (LOH) map with SNP array hybridization, global tumor LOH is associated with poor prognosis. The proposed studies take advantage of a well-established and well-characterized cohort, an associated extensive tumor bank, and the latest advances in molecular genetics for the study of lung cancer survivorship.